Regression forecasting of patient admission data

Abstract

Forecasting is an important aid in many areas of hospital management, including elective surgery scheduling, bed management, and staff resourcing. This paper describes our work in analyzing patient admission data and forecasting this data using regression techniques. Five years of Emergency Department admissions data were obtained from two hospitals with different demographic techniques. Forecasts made from regression models were compared with observed admission data over a 6-month horizon. The best method was linear regression using 11 dummy variables to model monthly variation (MAPE=1.79%). Similar performance was achieved with a 2-year average, supporting further investigation at finer time scales.